JP2008504546A - Measuring method of current flowing through a plurality of conductors, its application and apparatus - Google Patents

Abstract

A method for measuring a current flowing through a plurality (n) of conductors. In accordance with the present invention, the method includes placing a current transducer substantially opposite each conductor (i, i = 1,..., N) and a decorrelation that is a function of the transducer position relative to each conductor. Constructing a matrix ([G]) and measuring the current (I _measi ) through each conductor (i) with a current transducer, and calculating the uncorrelated matrix ([G]) and the current measurement (I _measi ) And estimating an actual current (I _reali ). Application is reversible electrical machines in the automotive industry.

Description

  The present invention relates to a method for measuring the current flowing through a plurality of conductors, an apparatus for implementing such a method, and the application of this method.

  The invention is particularly advantageous when applied to applications in the field of rotating electrical machines used in the automotive industry.

  An automobile equipped with a heat engine can be provided with a reversible electric machine, also called an alternator / starter. At start-up, the machine operates in both alternator and motor modes, i.e. as an aid to boost the heat engine from 500 revolutions per minute.

  The reversible electric machine includes a power device and a control device. The power device operates as a current inverter in the starter mode, operates as a current rectifier device in the alternator mode, and is controlled by the control device.

  In this type of machine, it is necessary to be able to constantly control the torque that is increased or decreased by the alternator / starter. However, this torque depends directly on the machine stator current, and more precisely on the current in each phase when the stator is operating with multiphase (eg, three phase) current. Therefore, a current control device mainly of numerical control type is provided for monitoring and controlling various stator currents.

  In this machine, the current passes through a conductor with a large cross section, which is arranged between the stator of the power plant and the rectifier (inverter).

  The conductor is, for example, a parallel straight conductor also called a bus bar (busbar).

  Therefore, in order to control the alternator / starter so that it is most appropriate for the operation of the automobile engine, the current through n busbars (n = 3 for a three-phase machine) must be strictly recognized. I understand that is advantageous.

  Various known types of devices are used to measure the stator current.

  FIG. 1 is a side view of a known measuring device using a magnetic circuit CM made of ferrite. The magnetic circuit CM surrounds the conductor CO and covers the Hall effect sensor CA. Hall effect sensor CA measures the magnetic field defined by current I through conductor CO and faces the magnetic circuit. Such a device is fixed to the heat sink of the alternator / starter power unit.

  However, such devices are expensive and large, and require first a connection link between the sensor and the control device (usually called the control card), and then each conductor of the bus bar to a ferrite magnetic circuit. Since it must be passed, it is complicated to use.

  Shunt based devices are also known, but this is not well suited for applications that measure very large currents (eg, 800 A) with little loss due to the Joule effect. There is also a problem of connection with the conductor. Finally, when the shunt value is small, the measurement of small currents is significantly inaccurate.

  Therefore, the technical problem to be solved by the present invention is an apparatus and method for measuring current flowing through a plurality of conductors, which is not expensive, lossless, easy to use, and guarantees precise measurement of a target current. Is to propose.

The solution to the proposed technical problem is in accordance with the present invention,
Placing a current transducer approximately opposite each fixed conductor;
Constructing an uncorrelated matrix that is a function of transducer position relative to the conductor;
Measuring the current in each conductor using a current transducer, and using the uncorrelated matrix and each measured current value to estimate an actual current therefrom.

  The method according to the invention only requires the use of a simple magnetic field transducer without a magnetic circuit, which transducer is a compact magnetic field sensor (preferably a Hall effect sensor or the like). Energy dissipation due to the Joule effect cannot occur, the measurement of the current in each conductor is exact and reproducible, and the decorrelation matrix is a fixed parameter that is explicitly determined by the placement of the transducer relative to that conductor Use only.

  According to the present invention, the elements included in the uncorrelated matrix continuously apply a calibration current in each conductor (the current applied to the other conductors is zero), and use the transducer to apply to each conductor. Determined by measuring the corresponding current signal. Thereafter, the actual current is estimated from each measured current value by applying an inverse matrix of the uncorrelated matrix.

  When there is an offset in the current, particularly due to inaccuracies in the sensor and the circuit that measures the current, the method according to the invention further provides a current in each conductor in the absence of an applied current in the conductor. Including determining an offset matrix having elements equal to each measured current value, using the uncorrelated matrix and said offset matrix to estimate the actual current from each measured current value; The accuracy of measurement can be further increased.

  Next, an actual current matrix is obtained by subtracting an offset matrix from the measured current matrix and applying an inverse matrix of the uncorrelated matrix to the obtained result.

  When the method which is the object of the present invention is applied to the measurement of the current in the input / output conductors of the stator poles of a rotating electrical machine, the output current of the stator is controlled in order to achieve numerical control of the current of the rotating electrical machine. An inverse projection matrix is provided that is multiplied by the inverse of the uncorrelated matrix to generate a single matrix that is applied to each measurement.

  The advantage of this last provision is that the numerical control processor can measure current (inverse of uncorrelated matrix) and change from an n-phase reference frame to a two-phase reference frame by means of a single matrix To include both the transformation (reverse projection matrix).

  A clear understanding of the structure and method of the present invention will be gained from the following description, given by way of non-limiting example with reference to the accompanying drawings, in which:

  FIG. 2 shows a device intended to measure the current flowing through n conductors, each marked with the letter i.

This device
a sensor C _i intended to measure the magnetic field corresponding to each of the n conductors;
A circuit MES for measuring the voltage corresponding to each current through the conductor i (this circuit determines the voltage measured at the terminals of the sensor C _i from the first reference (eg −10V, + 10V) to the second reference (E.g., intended to replace 0V-5V),

  A measurement management microcontroller MC intended to control the stator current (the microcontroller outputs the replaced measurement values (0V-5V) from the measurement circuit MES, 256, 512, or Including an analog-to-digital converter CAN adapted to selectively convert to 8, 10 or 12 bit digital units corresponding to 1024 points).

  In the case shown in FIG. 2, the number of conductors is 3 (n = 3). This situation can occur, for example, when the current flow through a conductor connected to the poles of the stator of a three-phase reversible electric machine (also called alternator / starter) is illustrated in the illustrated device and the method implemented thereby. This applies to measurement. The conductor i is, for example, a stator bus bar. They may be cables, rods or any other type of preferably rigid current conductor. With rigidity, in particular, a stable, constant, one-time uncorrelated matrix (matrix described in detail later) can be obtained.

  The measuring method corresponding to the device of FIG. 2 includes the following steps.

In the first step, as shown in FIG. 3, the current transducer C _i is disposed almost opposite to each conductor i (i = 1, 2, 3). In particular, the transducer C _i is a magnetic field measuring sensor such as a Hall effect sensor. The advantage of the Hall effect sensor is that a wide range of magnetic field values can be measured with high accuracy. In addition, from current measurements, it is possible to indicate a quantity (ie, for example, voltage, frequency, or current) that is proportional to the actual current measurement and thus represents the actual current measurement.

Each conductor i (ie, bus bar) is preferably disposed on the power card PCB_P. Each transducer (sensor) C _i is advantageously arranged on the control card PCB_C approximately opposite each conductor i. Of course, the sensor Ci is arranged so as not to be saturated beyond the measurement range.

  Therefore, since the sensor is not directly on the heat dissipation plate of the power card but is directly disposed on the control card, it is not necessary to connect the sensor from the heat dissipation plate to the control card. The advantage of this is that the problem of reliability due to the bulky and expensive connection as well as the problem of mechanical stress are eliminated. Mechanical stress is often due to vibrations (eg, arising from a car).

Similarly, unlike the prior art of FIG. 1, since the transducer C _i is not placed in the ferrite magnetic circuit, the magnetic field circulation in each conductor i is not routed through the magnetic circuit. As a result, an edge effect can occur in that the sensitivity of the corresponding conductor i, located on the opposite side of the transducer C _i , to the magnetic field of the other two conductors i is comparable. is there. This problem is solved by the following method.

Second, during the initialization step, a calibration current I _j ⁰ is applied to conductor j (no current is applied to the other conductor i (i ≠ j)), and each transducer C _i responds. The current signal I _i is measured. The measured equivalent current I _i is not zero for conductor i (i ≠ j). This is because the transducer C _i (i ≠ j) detects the magnetic field produced by the current I _j ⁰ in the conductor j and thus provides a current signal corresponding to this magnetic field. For example, it is possible to obtain a calibration current I _j ⁰ corresponding to the maximum current (for example, 1000 amperes) that the inverter can withstand.

In another example, the calibration current value I _j ⁰ for operating the management microcontroller MC can take the form of a power of 2, for example. In the case of a power of 2, for example, if the resolution is 0.1 amperes, a value 819.2 amperes corresponding to 8192 = 2 ¹³ can be obtained. Thus, this facilitates subsequent division calculations because at this point the microcontroller is only producing an offset.

Of course, as described above, the current signal I _i is an amount that can be expressed in terms of current, voltage, frequency, and the like in accordance with the sensor C _i , so that the current signal I _i is not applied to other conductors Shows the actual measured current value.

The expression G _ij = I _i / I _j ⁰ [1] estimates n elements G _ij of the uncorrelated matrix [G] (j is fixed and i is variable between 1 and n). ). When there are three conductors, the current I ₁ ⁰ is applied to the conductor i = 1, the currents I ₁ , I ₂ , and I ₃ are measured, and the elements G ₁₁ , G ₂₁ , and G ₃₁ are estimated.
I ₁ = G ₁₁ · I ₁ ⁰
I ₂ = G ₂₁ · I ₁ ⁰
I ₃ = G ₃₁ · I ₁ ⁰
By applying a calibration current to each conductor and performing this calculation n times, n ² elements G _ij of the matrix [G] are estimated.

  When there are three conductors, the matrix [G] is as follows.

This uncorrelated matrix [G] and its inverse [G] ⁻¹ are calculated by the microcontroller MC and stored in one of its memories (eg, rewritable EEPROM) (not shown).

In the normal mode of operation in the case of a motor vehicle, the current I _{measi in} each conductor i is measured by a current transducer C _i , and from the result, the determinant [I _real ] = [G] ⁻¹ [I _meas ] Current value _Irealj is estimated.

  Thus, the matrix [G] is inherently geometric in nature and takes into account, among other things, possible tolerances when mounting the sensor and variations in the distance between the bus bar and the sensor.

Therefore, this uncorrelated matrix [G] limits the influence of currents other than the current to be measured and makes it possible to limit the influence of the magnetic field measured by the sensor C _i .

  When applied to the currents of the three phases u, v, w of the reversible electric machine stator, the actual current is obtained from the current measurement by the following equation:

好ましくは、この電流測定値計算方法は、追加の較正ステップを含み、このステップは、特に以下のものによって生ずる測定の不正確さに起因するオフセットを考慮に入れることを可能にする。
センサＣ_i。
測定回路ＭＥＳの部品。
マイクロコントローラＭＣのアナログデジタル変換器ＣＡＮの部品。

Preferably, the current measurement calculation method includes an additional calibration step, which makes it possible to take into account offsets due to measurement inaccuracies caused in particular by:
Sensor C _i .
Parts of measurement circuit MES.
Parts of the analog / digital converter CAN of the microcontroller MC.

For example, when converting the measurement value into 12 bits for the measurement range of ± 1000 A, the converter CAN has an accuracy within 1 bit corresponding to ± 0.5 A (2000/2 ¹² = 0.5). ). If the measurement range is ± 100 A, the converter CAN has an accuracy within 1 bit corresponding to ± 0.5 A when converting the measurement value to 12 bits.

Therefore, the offset is considered by the offset matrix [O]. The element O _i of [O] is equal to the current value measured in each conductor i when no current is supplied to the conductor, so the actual current matrix [I _real ] is [I _real ] = [ G] ⁻¹ ([I _meas ] − [O]).

The element O _i can also be a numerical value corresponding to the current measurement.

  When there are three conductors, the offset matrix [O] is as follows.

  This matrix [O] is also calculated by the microcontroller MC and stored in one of its memories (eg, rewritable EEPROM) (not shown).

When considering this calibration step, the uncorrelated matrix [G] is calculated according to the following equation.
G _ij = (I _i −O _i ) / I _j ⁰ [2]. O _i corresponds to the offset of the current values I _j , i (i ≠ j) measured during the initial initialization step.

  The matrices [G] and [O] can include various gains in the chain and offset, respectively, which allows the signal to be repositioned to a median value corresponding to zero. Please be careful.

Therefore, when only the offset due to the sensor is taken into account, the uncorrelated matrix [G] is expressed in ohms and the inverse matrix [G] ⁻¹ is expressed in siemens. An example of such a matrix is shown below. However, each sensor C _i is held at a distance of about 2.5 cm from the vertical line of the corresponding bus bar i, and the first sensor C ₁ has a diagonal distance of 5.cm from the second bus bar. Located at a location of 5 cm and a diagonal distance of 10.5 cm with respect to the third bus bar.

ただし、対応するオフセット行列［Ｏ］は、次式と等しい。

However, the corresponding offset matrix [O] is equal to:

当然ながら、電流測定装置全体（すなわち、センサ、測定回路、およびアナログデジタル変換器）は較正される。その場合、無相関行列［Ｇ］の単位は、アンペアマイクロコントローラ単位になる。

Of course, the entire current measurement device (ie, sensor, measurement circuit, and analog-to-digital converter) is calibrated. In this case, the unit of the uncorrelated matrix [G] is an ampere microcontroller unit.

Thus, the actual current I _real has different units depending on the calibration. For example, the unit, when the calibration is related only to the sensor C _i is can be a voltage, in the case where the calibration is related to the sensor C _i and the measurement circuit MES is can be a frequency And can be a numerical value with resolution if the calibration is for example related to the sensor C _i , the measuring circuit MES and the analog-digital converter MC.

  Thus, the uncorrelated matrix [G] (also called the gain matrix) and the offset matrix [O] (also called the offset matrix) solve the problem of sensor correlation and inaccuracy due to the components of the entire measuring device. By doing so, it is possible to calibrate all of the current measuring devices. This is impossible when only a simple uncorrelated vector or an offset vector is used.

FIG. 4 shows that the present invention can be applied to numerical control of the three-phase currents I _u , I _v , and I _w of the stator of the reversible machine.

This current measurement method includes, in a non-limiting embodiment, the additional step of converting the actual current measurement I _real of the n-phase reference frame into a two-phase reference frame by a projection matrix [C].

  This additional step is performed in order to simplify the calculation of the current in the phase φ of the alternator / starter machine stator and the current control method.

When n = 3, the matrix used to process the current is a three-phase current I _u , I _v , and a matrix known by the name of the Concordia matrix or another matrix known by the name of the Clark matrix. the I _w, a matrix for transforming the two-phase currents I _alpha and I _beta. In order to adjust the stator current, these two-phase currents are continuously applied to the numerical controller. In this way, the following equation is obtained.

すなわち、［Ｉ_regul］＝［Ｃ］^-1［Ｉ_real］である（Ｉ_hは単極成分）。

That is, [I _regul ] = [C] ⁻¹ [I _real ] (I _h is a monopolar component).

Unipolar components, it should be noted that corresponding to the sum of three-phase current in the vertical third axis to the axis O _beta.

  When the three-phase winding has a delta shape that is well known to those skilled in the art, the unipolar component corresponds to the internal circulation of the current.

  If the three-phase winding has a star shape that is well known to those skilled in the art, the monopolar component corresponds to the neutral point of the stator, which is a common point between the three phases. When the neutral point is not connected, the monopolar component is neutral.

This matrix [C] and its inverse [C] ⁻¹ are stored in one (not shown) of the memory (eg, non-rewritable ROM or rewritable EEPROM) of the microcontroller MC.

FIG. 5 shows the three-phase currents I _u , I _v , and (α, β, O) in the system (α, β, O) axis of the first phase φ _u corresponding to the current I _measu to mark the angle It shows the projection of I _w.

According to an example of an applicable Concordia projection matrix, [C] and [C] ⁻¹ are as follows:

および

and

According to an example of an applicable Clark projection matrix, [C] and [C] ⁻¹ are as follows:

および

and

これらの射影行列の係数は、一定であるが、３相電流の回転方向、それらの電流の強度などのような規則の関数であることに留意されたい。したがって、様々な標準化係数を有することが可能である。

Note that the coefficients of these projection matrices are constant, but are a function of rules such as the direction of rotation of the three-phase current, the strength of those currents, and so on. Therefore, it is possible to have various standardization factors.

Therefore, with the apparatus of FIG. 2, the two-phase currents I _α and I _β are measured using the single matrix [M] = [C] ⁻¹ · [G] ⁻¹ = [G · C] ^−1. It can be seen from I _meas that it can be obtained directly.

Therefore, the following equation holds.
[I _{α, β} ] = [M] · ([I _meas ] − [O])

Note that the product of the two matrices [G] ⁻¹ and [C] ⁻¹ is done off-line in that the rotation of the machine and thus the angle θ between the stator and the rotor is not taken into account.

If movement within the machine's reference frame is required, i.e. considering the angle θ between the stator and the rotor, and therefore when calculating the current online (in real time), the axes O _α and O _β are Assume that the first phase φ _u is offset by an angle θ. The new axes are a straight axis O _d and a horizontal axis O _q that are well known to those skilled in the art. Therefore, by applying the following rotation matrix [R], the system (α, β, θ) is changed to the system (d, q, O).

したがって、次式が成り立つ。

Therefore, the following equation holds.

すなわち、

That is,

  A transformation known by the name Park [P], which is the product of the projection matrix (Concordia or Clark) and the rotation matrix [R] is used.

Thus, by applying the inverse Park matrix [P] ⁻¹ = [R] ⁻¹ · [C] ⁻¹ to the current measurement I _meas (if applicable, the offset matrix [O] New current I _dq is obtained. In this way, [I _dq ] = [P] ⁻¹ · [G] ⁻¹ · ([I _meas ] − [O])
= [R] ^-1 · [C] ^-1 · [G] ^-1 · ([I _meas ]-[O])
= [R] ⁻¹ • [M] • ([I _meas ]-[O])
Is obtained.

  This Park transformation makes it possible to control the machine current more efficiently. In this way, continuous amounts are obtained that are easy to adjust rather than variable or alternative amounts.

  Note that before normal operation of the inverter and rectifier, the matrices [G], [O], and [M] are calculated only once by the microcontroller MC.

  Therefore, this current measurement method has certain advantages over the theoretical calculation of an uncorrelated matrix performed by a computer. The latter does not recognize interference due to, for example, components adjacent to the sensor, as well as inaccuracies due to the sensor. Furthermore, the method of the present invention is simpler than computer calculations. Finally, the method according to the invention allows the current measurements in the various conductors to be completely uncorrelated. Note that this decorrelation is different from the decorrelation between the magnetic flux and current measured in the conductor (which is trying to achieve a different goal than the current decorrelation).

  The method according to the invention is not only applicable to current measurements in rotating electrical machines, but also there is a correlation between the measured values, for example a battery management system (usually called BMS) in an automobile or a DC / DC converter. It should be noted that it is applicable to any application that requires measurement of multiple currents in a limited space that occurs and is therefore uncorrelated. This is because a housing generally included in a battery management system includes various connectors to which a battery and a consumer such as an air conditioning system or a hi-fi management system are connected. In order for the battery and consumer to operate properly, it is necessary to measure the current through them.

  Similarly, a DC / DC voltage converter includes various cells or components through which current flows, and input and output currents that need to be measured. It is noted that the DC / DC converter can be used in a 42V car with a battery of 42V and a consumer of 12V, in which case the converter allows a 42V to 12V conversion. I want.

It is a figure which shows the current measuring apparatus by a prior art. 1 shows a current measuring device implementing a method according to the invention. FIG. 3 is a partial side view of the apparatus of FIG. 2. FIG. 3 shows the application of the method according to the invention to the measurement of the current flowing through the input / output conductors of the stator poles of a rotating electrical machine. FIG. 5 is a diagram illustrating a projection of a three-phase current in a two-phase current system according to an embodiment of the method of FIG.

Explanation of symbols

CM magnetic circuit CO conductor CA Hall effect sensor C ₁ sensor (current transducer)
MES Voltage measurement circuit MC Measurement management microcontroller CAN Analog to digital converter PCB_P Power card PCB_C Control card

Claims (13)

A method for measuring a current flowing through a plurality (n) of conductors,
Placing a current transducer (C _i ) substantially opposite the fixed conductor (I, I = 1,..., N);
Constructing a decorrelation matrix ([G]) that is a function of the position of the transducer (C _i ) relative to the conductor;
The current (I _meas ) in each conductor (i) is measured using the current transducer (C _i ), and the uncorrelated matrix ([G]) and the measured current values (I _measi ) are used. And then estimating an actual current (I _reali ). The element (G _ij ) of the decorrelation matrix ([G]) continuously applies the calibration current (I _j ⁰ ) in each conductor (j) (the current applied to the other conductors is zero). Method according to claim 1, characterized in that it is determined by measuring the current signal (I _i ) corresponding to each conductor (i) using the transducer (C _i ). The actual current (I _real ) is estimated from the measured current (I _meas ) by applying an inverse matrix ([G] ⁻¹ ) of the uncorrelated matrix ([G]). The method according to claim 2. Determining an offset matrix ([O]) having an element (O _i ) equal to the current measured in each conductor (i) when no current is supplied in the conductor, The current ( _Ireal ) is estimated from the measured current according to the uncorrelated matrix ([G]) and the offset matrix ([O]), according to any one of claims 1-3. Method. The actual current (I _real ) matrix ([I _real ]) is obtained by subtracting the offset matrix ([O]) from the current measured value (I _meas ) matrix ([I _meas ]). 5. The method according to claim 4, characterized in that it is obtained by applying an inverse matrix ([G] ⁻¹ ) of the uncorrelated matrix ([G]) Method according to any of the preceding claims, characterized in that the current transducer (C _i ) is a Hall effect sensor. 7. A method according to any one of the preceding claims, characterized in that each comprises a plurality (n) of current transducers (C _i ) arranged substantially opposite each conductor (i). Device for carrying out. Device according to claim 7, characterized in that the current transducer (C _i ) is a Hall effect sensor.   Application of the method according to claim 1 to the measurement of the current flowing through the input / output conductors of the stator poles of a rotating electrical machine. To generate a single matrix ([M] = [C] ⁻¹ · [G] ⁻¹ ) applied to the stator output current measurement to perform numerical control of the rotating electrical machine current 10. Application according to claim 9, characterized in that the inverse projection matrix ([C] ^-1 ) is multiplied by an inverse matrix ([G] ^-1 ) of an uncorrelated matrix.   11. The offset matrix ([O]) is applied to the output current measurement value of the stator before applying the single matrix ([M]) to the current measurement value. Application of the description.   Application of the method according to any of claims 1 to 6 to the measurement of current flowing through a battery management system.   Application of the method according to any of claims 1 to 6 to the measurement of current flowing through a voltage converter system (DC / DC).

Images